I had a client tell me last week that they were thinking about canceling their SEO retainer. Their reason? "I can just use Claude to write all my content now. Why am I paying you for blog posts?"
Fair question. And honestly, they're not wrong that AI tools like Claude, ChatGPT, and Gemini can write a decent-sounding blog post in 30 seconds. I use these tools every day. They're incredible.
But here's what's actually happening to law firm websites that go all-in on AI content without a strategy behind it: they're getting crushed by Google updates.
I've seen it firsthand. And the data from our own client communication backs it up. The firms asking us for help right now aren't the ones who never tried AI. They're the ones who tried it, published everything it gave them, and watched their traffic drop.
The real problem isn't AI content. It's AI content without a plan.
Let me be specific about what I'm seeing.
A firm publishes 40 AI-generated blog posts in a month. They cover every keyword variation they can think of. "Car accident lawyer in [city]." "What to do after a truck accident in [state]." "How much is my personal injury case worth?" Times 40 cities. Times 3 practice areas.
Google's March 2025 core update wiped out sites doing exactly this. Not because AI wrote it. Because the content had no strategic purpose. No internal linking plan. No topical authority structure. No E-E-A-T signals. Just volume.
Volume without strategy is noise. And Google has gotten very good at identifying noise.
AI still hallucinates. Regularly.
This is the part that keeps me up at night when I think about law firms publishing AI content without review.
I asked Claude to write about personal injury statute of limitations by state last month. It got 3 states wrong. Not wildly wrong. Just wrong enough that a client reading it could miss their filing deadline.
ChatGPT is no better. I've seen it:
- Cite case law that doesn't exist (it generates realistic-sounding case names with fake citations)
- State incorrect filing deadlines
- Confuse jurisdiction-specific rules (mixing California law with Texas law in the same paragraph)
- Invent statistics that sound plausible but are completely fabricated
- Reference legal standards that were overturned years ago
For a restaurant blog, a hallucinated fact is embarrassing. For a law firm, it's a potential malpractice issue. And it undermines the trust your entire website is built on.
If someone reads a wrong statute of limitations on your site and misses their deadline, that's on you. Not on ChatGPT.
Every single piece of AI-generated legal content needs attorney review before publishing. And most firms skipping their agency's content process aren't doing that review. They're just hitting publish.
What Google actually penalizes (and it's not "AI content")
Google has said publicly they don't penalize content just because AI wrote it. They penalize content that's:
- Thin. Saying the same thing 50 different ways across 50 pages.
- Duplicative. Multiple pages targeting the same intent with slightly different titles.
- Lacking E-E-A-T. No real experience, expertise, authoritativeness, or trust signals. AI can't show it tried cases. You can.
- Not helpful. Content that exists to rank, not to answer a real question.
- Mass-produced without quality controls. Google's spam policies specifically call out "scaled content abuse" - using automation to generate large amounts of content for search ranking purposes.
Here's the pattern I keep seeing. A firm publishes 200 pages in 3 months. Rankings spike for a few weeks. Then the next core update hits. Traffic drops 40-60%. They call us asking what happened.
What happened is that Google's systems caught up. The initial spike was because Google indexes new content and gives it a chance. The drop is when the quality signals come back and Google realizes 180 of those 200 pages are the same page rewritten slightly differently.
The 5 ways AI content without strategy actually hurts your site
1. Keyword cannibalization. AI doesn't know what your other pages target. It writes about "car accident lawyer" on 12 different pages because you asked it 12 different prompts. Now those pages compete against each other. None of them rank well.
2. Topical dilution. You're a family law firm but you published 30 posts about personal injury because AI made it easy. Google is now confused about what your site is actually about. Your core practice area rankings drop.
3. No internal linking structure. AI writes standalone pages. It doesn't know about your other content. A good SEO strategy connects pages in clusters - a pillar page linking to supporting content, building authority around a topic. AI just writes a page and moves on.
4. Thin content accumulation. 50 pages that each say "if you've been in a car accident, call a lawyer" in slightly different words. Google sees this as one page of value stretched across 50 URLs. That's worse than having no content at all.
5. Trust erosion. When a potential client reads a blog post with wrong information - or worse, when opposing counsel finds it - your credibility takes a hit that no amount of content volume can fix.
What AI content done right looks like
I'm not anti-AI. We use AI tools at Juris Digital every week. But there's a difference between using AI as a tool within a strategy and using AI as the strategy.
Here's what works:
Start with keyword research, not a prompt. Before anyone writes anything - human or AI - you need to know what your site should rank for, what it already ranks for, and where the gaps are. AI can't do this analysis for you.
Build a topical map first. Decide your pillar topics. Map out the supporting content. Plan the internal links. Then write the content to fit that map. This is strategy. This is what moves rankings.
Use AI for first drafts, not final drafts. AI is great at getting ideas on paper fast. It's terrible at fact-checking itself, adding your firm's actual case experience, and matching your jurisdiction's specific laws. A human attorney needs to review every piece.
Add real E-E-A-T. AI can't add that you won a $2.3M verdict in a trucking case last year. It can't reference the specific judge you'll appear before in Harris County. It can't share what you learned from 15 years of trying DUI cases. That experience is what Google wants to see, and it's what separates your content from the 10,000 other "what to do after a car accident" posts.
Publish less, publish better. 4 excellent pages per month that are strategically linked, attorney-reviewed, jurisdiction-specific, and supported by real experience will outperform 40 generic AI posts every time. Every. Single. Time.
The firms that will win in 2026 and beyond
My prediction is that AI content will become table stakes. Everyone will have it. The firms that win will be the ones who pair AI efficiency with human strategy.
That means:
- An SEO strategy that determines what to write before anything gets written
- Attorney review on every piece of published content
- Real experience woven into the content (case results, client stories, jurisdiction-specific knowledge)
- Technical SEO that makes sure the content actually gets indexed and ranked (schema markup, proper site architecture, page speed)
- Regular audits to prune content that isn't performing instead of just adding more
The firms that dump 200 AI blog posts onto their site without a plan will continue to get hit by every core update. The firms that use AI as one tool within a real strategy will pull ahead.
What I'd do this week if I were running a law firm
-
Audit your recent content. If you've published AI-generated content in the last 6 months, check Google Search Console. Are those pages getting impressions? Clicks? Or are they just sitting there doing nothing?
-
Check for cannibalization. Search your own site (site:yourfirm.com "car accident lawyer") and see if you have multiple pages competing for the same query. If you do, consolidate them.
-
Review for accuracy. Pick 5 AI-generated posts at random and have an attorney in your firm fact-check them. Check statute of limitations, filing deadlines, specific legal standards, and case citations. I'd bet money at least one has an error.
-
Ask your SEO agency this question: "What's the strategy behind the content you're publishing for us?" If they can't explain the keyword targeting, the topical authority plan, and the internal linking structure - that's a problem. If they can, that's the value you're paying for. Not the words on the page.
The words are cheap now. Everyone has access to AI. The strategy behind those words is what separates firms that rank from firms that don't.
That's not going to change anytime soon.